4642

Zoomed 3D GRE EPI utilizing a Segmented 2D Selective RF Pulse Excitation
Naoharu Kobayashi1, Michael Mullen1,2, Kamil Ugurbil1, and Michael Garwood1

1Center for Magnetic Resonance Research, Department of Radiology, University of Minnesota, Minneapolis, MN, United States, 2School of Physics and Astronomy, University of Minnesota, Minneapolis, MN, United States

Synopsis

Segmented 2D selective RF pulse excitation is introduced in 3D-EPI for zoomed imaging. The feasibility of the 2D selective pulse segmentation was tested in phantom and in vivo brain measurements. The segmented pulse excitation provided nearly identical excitation profile to a non-segmented pulse excitation. In zoomed in vivo brain imaging, the segmented pulse excitation showed conspicuous improvement of susceptibility artifacts around the frontal sinus.

Purpose

Three-dimensional echo planar imaging (3D-EPI) can achieve fast high-resolution imaging while preserving an increased SNR as compared to 2D-EPI. However, high-resolution imaging in EPI results in increased off-resonance/susceptibility artifacts due to the longer readout echo train length and/or increased readout segments. Previously, 2D selective RF excitation was introduced in EPI to limit the imaging volume along the phase encoding direction and to reduce susceptibility artifacts in regions close to air cavities (1). While it improves the image quality, one general limitation of 2D selective RF pulses is the lengthy pulse width, which makes excitation extremely sensitive to off-resonance especially around the strong susceptibility regions. In this study, we introduce segmented excitation of a 2D selective frequency-modulated pulse in 3D-EPI. Pulse segmentation increases the excitation bandwidth by a factor of the number of segments and thus reduces the impacts from off-resonance. The feasibility of 2D selective pulse segmentation was demonstrated with phantom and in vivo brain imaging.

Methods

A 2D selective hyperbolic secant (2D-HS) pulse was designed in the excitation k-space domain with a raster k-space trajectory (Fig.1A)(2). The 2D-HS pulse was composed of 20 HS subpulses with an HS pulse envelope in time domain (Fig.1B). Each subpulse, which traverses along the fast dimension (x axis) in k-space, had a pulse width of 0.8 ms and a time bandwidth product (TBP) of 16, and TBP along the slow/blip gradient dimension (y axis) was 10. The 2D-HS pulse was implemented as an excitation pulse in the segmented 3D GRE EPI sequence, where two phase encoding dimensions were assigned to the selective dimensions of the 2D-HS pulse (Fig.1C). Pulse segmentation was tested for three cases: 1, 2 and 4 segments (Fig.2). Duration of the entire 2D-HS pulse was 18.4 ms for the single segment pulse.

MR imaging in this study was performed with a 3T Siemens Prisma scanner using a 32 channel head coil. To test the excitation profile of the segmented pulses, phantom measurements were performed with following parameters: TR/TE=67/26 ms, flip angle=10°, FOV=192x186x72 mm3, 1.5 mm isotropic resolution and scan time=12.9/25.8/51.5 sec for 1/2/4 segment excitation. Brain imaging was conducted on a healthy volunteer under an IRB approved protocol. Zoomed imaging was performed by selecting FOV to the excited region of 192x93x36 mm3, which resulted in scan time of 3.3/6.5/12.9 sec for 1/2/4 segment excitation. TE was set to the minimum in each excitation: TE=26/21/19 ms for 1/2/4 segment excitation. Image reconstruction was performed offline with a reconstruction routine running on Matlab.

Results

The single segment pulse accomplished a clean 2D excitation profile without sideband excitation (Fig.3A). Pulse segmentation generated sideband excitation along the slow dimension and the sidebands got close to the baseband as the number of segments increase (Fig.3B,C). There was a severe overlap of the baseband and sidebands observed for the 4 segment pulse. However, the resultant excitation of the segmented pulses provided similar profiles to the single segment pulse including the quadratic phase (Fig.3D). A consistent 2D excitation profile was also obtained for in vivo brain imaging (Fig.4). However, the four segment pulse created slightly different image contrast from the other two pulses. In zoomed imaging of the brain, the segmented pulse excitation showed conspicuous improvement of susceptibility artifacts compared to the single segment pulse as the number of pulse segments increased (Fig.5).

Discussion

The segmented 2D pulses provided almost identical excitation to the single shot pulse excitation. Under the low tip angle approximation, the spin system can be assumed to be linear (3). The flip angle of 10° in this study is reasonably low to preserve the linearity of the spin system. With the four segment excitation, interference of the baseband and sideband excitation made the flip angle different segment by segment, which made T1-weighting different (i.e., slightly different image contrast).

One drawback of using the segmented 2D pulse is increasing scan time, since multiple excitations are needed to obtain a clean baseband excitation profile. In this study, accelerated data acquisition such as parallel imaging and compressed sensing reconstruction was not employed. However, acquisition can be readily accelerated by 2−4 folds along the two phase encoding dimensions in 3D-EPI.

Conclusions

The segmented 2D selective RF pulse in 3D-EPI improved the susceptibility artifacts in strong susceptibility regions by reducing the pulse width in each excitation (i.e., increasing excitation bandwidth and shortened achievable echo time).

Acknowledgements

This study was supported by NIH grant P41EB015894, U01EB025153-01, 1S10OD017974-01 and T32EB008389.

References

1. Rieseberg S, Frahm J, Finsterbusch J. Two-dimensional spatially-selective RF excitation pulses in echo-planar imaging. Magnetic resonance in medicine. 2002;47(6):1186-93.

2. Mullen M, Kobayashi N, Garwood M. 2D Selective Excitation with Resilience to Large B0 Inhomogeneities. Proceedings of the European Magnetic Resonance Meeting. 2018.

3. Pauly J, Nishimura D, Macovski A. A k-space analysis of small-tip-angle excitation. 1989. J Magn Reson. 2011;213(2):544-57.

Figures

Figure 1. 2D selective hyperbolic secant (HS) pulse in excitation k-space (A) and time domain (B). The excitation k-space was sampled along a raster trajectory, providing repeated HS subpulses with an HS pulse envelope in time domain. The k-space sampling is fast along kx (fast dimension), whereas it is slow along ky (slow dimension) in time. C) 3D segmented EPI sequence with 2D selective HS pulse excitation. The sequence was composed of three modules: fat saturation, 2D-HS pulse excitation and 3D segmented EPI readout. The 2D selective dimensions in excitation (x and y) were selected to phase encoding dimensions in readout.

Figure 2. Segmentation of a 2D selective HS pulse with a raster trajectory. The pulse was segmented along the slow dimension (y axis). Pulse segmentation can reduce the pulse width in each excitation by a factor of the number of segments, which increases excitation bandwidth along the slow dimension.

Figure 3. Phantom excitation profile of a 2D HS pulse with three different segmentations: A) single, B) two and C) four segment pulses. The single segment pulse accomplished a clean 2D excitation profile without sideband excitation. Pulse segmentation created sideband excitation along the slow dimension (y). Interference of the baseband and sidebands is minimal for the 2 segment pulse, whereas a severe overlap of the baseband and sidebands was observed in the 4 segment pulse. Nonetheless, the resultant excitation provided similar profiles to the single segment pulse including the quadratic phase (D, a dashed yellow line in A).

Figure 4. In vivo brain excitation profile of a 2D HS pulse with three pulse segmentations: single, two and four segment pulses. All pulse segmentations provided a consistent 2D excitation profile. While the single and two segment pulses generated similar image contrasts, there is a slightly different contrast observed for the four segment pulse (e.g. CSF and brain tissues).

Figure 5. Zoomed 3D EPI images of in vivo brain. A strong susceptibility region above the frontal sinus (arrow heads) clearly showed an improvement of signal voids as the number of pulse segments increased.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
4642